Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Jul 3.
Published in final edited form as: Environ Res. 2019 Jul 22;180:108604. doi: 10.1016/j.envres.2019.108604

Relation between in utero arsenic exposure and growth during the first year of life in a New Hampshire pregnancy cohort

Meghan E Muse a,∗, Zhigang Li b, Emily R Baker c,d, Kathryn L Cottingham c,e, Susan A Korrick c,f,g, Margaret R Karagas a,c, Diane Gilbert-Diamond a,c
PMCID: PMC7333643  NIHMSID: NIHMS1598299  PMID: 31710845

Abstract

Background

We have previously reported that in utero arsenic exposure is associated with increased length and other anthropometric outcomes at birth in a U.S. cohort. However, it is unknown whether these anthropometric differences persist through early life.

Objectives

We assessed in utero arsenic exposure in relation to attained anthropometry and growth trajectories through the first year of life.

Methods

Among 760 mother-infant pairs from the New Hampshire Birth Cohort Study, we assessed in utero arsenic exposure using maternal second trimester urinary arsenic and assessed infant growth from medical records.

Results

Median maternal second trimester total urinary arsenic (tAs; inorganic arsenic + monomethylarsonic acid + dimethylarsinic acid) was 3.96 μg/L (IQR: 2.02, 6.72). In adjusted linear mixed effects models, each doubling of maternal urinary tAs was associated with a 0.05 increase in length WHO Z score (95% CI: 0, 0.09) over the first year of life which corresponds to an approximately 0.12 cm increase in males and 0.13 cm increase in females at 12 months. No associations were observed between urinary tAs and attained weight, weight-for-length, or head circumference. In adjusted piecewise linear mixed effects models, each doubling of urinary tAs was associated with a 0.07 (95% CI: 0.02, 0.12) cm per month decreased length growth rate through 3.5 months with no evidence of an association thereafter. No associations were observed between urinary tAs and infant weight gain or change in weight-for-length and head circumference through one year.

Conclusions

On average, infants exposed to higher in utero arsenic attained modestly longer length during the first year, despite having slower linear growth in the first 3.5 months of life. This suggests that the previously demonstrated arsenic-associated longer length among study infants at birth persists through the first year of life. No other anthropometric associations with in utero arsenic exposure were observed across the full study population.

Keywords: Arsenic, In utero exposure, Anthropometry, Infant growth

1. Introduction

Environmental exposures during the sensitive period of fetal development may have long-lasting effects on child health (Swanson et al., 2009). One environmental exposure of concern is arsenic, a naturally occurring metalloid that people are commonly exposed to through drinking water (Smedley and Kinniburgh, 2002) and consumption of rice and other foods (Gilbert-Diamondetal.,2011;Meharg and Rahman, 2003). It is estimated that at least 140 million individuals worldwide have been exposed to drinking water arsenic in excess of World Health Organization (WHO) guidelines of 10 μg/L (WHO, 2018). In New Hampshire, more than 10% of private wells are estimated to have water with arsenic concentrations in excess of these guidelines (Karagas et al., 1998).

Arsenic can be ingested in inorganic (iAs) and methylated forms. iAs is known to be highly toxic (Hughes, 2002), and has been associated with numerous health conditions including skin lesions (Ahsan et al., 2000; Haque et al., 2003), heart disease (Chen et al., 1996; James et al., 2015; Tseng et al., 2003), type 2 diabetes (Lai et al., 1994; Tseng et al., 2000), and some cancers (Hopenhayn-Rich et al., 1996, 1998; Smith et al., 1998). After ingestion, iAs goes through a series of methylation and reduction steps (Vahter, 2002). iAs can be methylated into monomethylarsonic acid (MMAV) which then can be reduced to monomethylarsonous acid (MMAIII). MMAIII subsequently can be further methylated into dimethylarsinic acid (DMAV) and then further be reduced to dimethylarsinous acid (DMAIII). The biotransformation of ingested iAs is incomplete and iAs, monomethylarsonic acid (MMAV), and dimethylarsinic acid (DMAV) are all excreted in urine (Francesconi et al., 2002; Watanabe and Hirano, 2013). In addition, MMAIII and DMAIII have also been found in the urine of people highly exposed to iAs (Mandal et al., 2001; Valenzuela et al., 2005), though not in the urine of people with low-level arsenic exposure (Lindberg et al., 2006; Rivera-Núñez et al., 2012). While the methylation of iAs to MMA and DMA is generally considered a detoxification process (Gebel, 2002; Moore et al., 1997), studies suggest that MMAIII and DMAIII may be more toxic than iAs with MMAIII being the most toxic arsenic species (Vahter, 2002). This biotransformation is thought to be increased in pregnant women, with one study in a native Andean population finding that, during pregnancy, on average greater than 85% of arsenic excreted in urine is in the form of DMA (Concha et al., 1998). Arsenic (including iAs and its methylated forms) crosses the placenta (Concha et al., 1998; Punshon et al., 2015), and thus biomarkers of maternal exposure to arsenic during pregnancy, such as urine, are thought to reflect in utero arsenic exposure (Vahter, 2009).

Several studies have examined the association between prenatal arsenic exposure and birth anthropometric outcomes (Claus Henn et al., 2016; Gilbert-Diamond et al., 2016; Hopenhayn et al., 2003; Kwok et al., 2006; Laine et al., 2015; Myers et al., 2010; Rahman et al., 2009; Yang et al., 2003). While most of these studies were conducted in populations with high average levels of inorganic arsenic in drinking water (e.g., > 50 μg/L) (Hopenhayn et al., 2003; Kwok et al., 2006; Laine et al., 2015; Myers et al., 2010; Rahman et al., 2009; Yang et al., 2003); Claus Henn et al. (2016), and our group (Gilbert-Diamond et al., 2016) examined how prenatal arsenic exposure related to anthropometry at birth in U.S. populations with predominately lower levels of inorganic arsenic in drinking water (e.g., < 10 μg/L). In both of these studies, prenatal arsenic exposure was related to smaller head circumference at birth. Results varied between the two U.S. studies for other birth anthropometric outcomes. For example, Claus Henn et al. identified a negative association between in utero arsenic exposure and birth weight across all infants while in our cohort this association was seen only in females born to overweight and obese mothers (Gilbert-Diamond et al., 2016). In our cohort, there was also a positive association between in utero arsenic exposure and length at birth and a negative association between in utero arsenic exposure and adiposity at birth (assessed as ponderal index) in the overall study population (Gilbert-Diamond et al., 2016).

In addition, three published papers have previously reported on the association between prenatal arsenic exposure and early childhood growth (Agay-Shay et al., 2015; Gardner et al., 2013; Saha et al., 2012). In the Maternal and Infant Nutrition Interventions in Matlab (MINIMat) trial in the arsenic endemic region of Matlab, Bangladesh, prenatal arsenic, assessed as maternal total urinary arsenic, was negatively associated with attained weight and length in female infants at 3–24 months, but those associations largely disappeared after multivariable adjustment (Saha et al., 2012). In the same cohort, no association was identified between prenatal arsenic and child anthropometry at 5 years (Gardner et al., 2013). Similarly, the Environment and Childhood Project (INMA) in Sabadell (Catalonia, Spain) did not find any associations between total prenatal urinary arsenic concentrations (which included arsenobetaine, a nontoxic, unmetabolized form of arsenic found in fish and seafood) and weight at 7 years of age (Agay-Shay et al., 2015).

The aim of this study was to determine whether prenatal arsenic exposure was associated with differences in child anthropometric indicators through the first year of life in our U.S. population with generally low levels of arsenic exposure. We further sought to study whether any associations were modified by child sex or maternal pre-pregnancy weight status due to previous research suggesting effect modification by these factors on the association between in utero arsenic exposure and anthropometry at birth (Gilbert-Diamond et al., 2016) and by child sex on the association between postnatal arsenic exposure and infant growth (Saha et al., 2012).

2. Methods

2.1. Source population

The New Hampshire Birth Cohort Study (NHBCS) is a prospective study that aims to examine the associations between environmental exposures and other factors and maternal-child health outcomes (Gilbert-Diamond et al., 2011). Beginning in January 2009, pregnant women between 24 and 28 weeks of gestation were recruited from prenatal clinics in New Hampshire. Eligibility criteria included age of 18–45 years old, English literacy, the use of a private, unregulated water system (e.g., private well) at home, not planning to move, and a singleton pregnancy. The study focused on New Hampshire residents using unregulated private water supplies, because it is estimated that more than 10% of homes in the state that use private wells have water arsenic concentrations that exceed the Environmental Protection Agency’s recommended maximum of 10 μg/L (Karagas et al., 1998). At the time of the analysis there were 1,758 mother-child dyads enrolled in the New Hampshire Birth Cohort Study. Of those, maternal second trimester spot urine samples had been analyzed for speciated arsenic exposure for 1,260 women. Anthropometric data were available from medical record review at 2 weeks, 1 month, 2 months, 4 months, 6 months, 9 months, and 1 year for 686, 362, 744, 722, 720, 691, and 691 children, respectively. In total, 760 mother-child dyads had both speciated arsenic concentrations measured in maternal second trimester spot urine samples and anthropometry data from birth and at least one postnatal well child visit up to age 1 year. Participants provided written informed consent and all study procedures were approved by the Institutional Review Board at Dartmouth College.

2.2. Primary outcomes: anthropometric measures

Infant head circumference (cm), weight (g), and length (cm) were abstracted by trained study staff from medical records at birth and well child visits from 2-weeks to 12-months of age. We used the World Health Organization (WHO) reference curves to calculate sex- and age- standardized head circumference, weight, length, and weight-for-length Z-scores (WHO Multicentre Growth Reference Study Group, 2006).

2.3. Primary exposure: In utero arsenic exposure

We collected and analyzed maternal second trimester urinary arsenic using previously described methods (Gilbert-Diamond et al., 2011). In brief, women provided a spot urine sample collected at approximately 24–28 weeks of pregnancy. Samples were analyzed for iAs (AsIII and AsV), MMAV, and DMAV using a high-performance liquid chromatography (HPLC) ICP-MS system (Larsen et al., 1993). The average limits of detection (LOD) for iAs, MMAV, and DMAV were 0.167, 0.065, and 0.061 μg/L, respectively. For each arsenic species, measurements below the detection limit were set to the LOD divided by the square root of two. There were 270 samples below the LOD for iAs, 177 samples below the LOD for MMA, 3 samples below the LOD for DMA, and 2 samples were below the LOD for all 3 arsenic species. We calculated total urinary arsenic concentrations (tAs) as the sum of iAs, MMA, and DMA. Arsenobetaine was also measured in all urine samples via HPLC but was not included in the calculation of total urinary arsenic as it is considered a nontoxic, unmetabolized arsenic species (Francesconi et al., 2002). To assess urinary dilution, we also measured urinary creatinine using a colorimetric assay (Assay #500701; Cayman Chemical, Ann Arbor, MI). At the time of this analysis, urinary creatinine had been measured for 646 of the 760 urine samples.

2.4. Maternal weight status

Maternal usual weight prior to pregnancy was assessed via self-report as part of the enrollment questionnaire. We then combined the reported weight with height abstracted from the medical records to compute pre-pregnancy BMI. Pre-pregnancy weight status was categorized as normal weight (18.5 ≤ BMI < 25 kg/m2), overweight (25 ≤ BMI < 30 kg/m2), or obese (≥30 kg/m2) (WHO, 2000). Due to previous research suggesting that underweight mothers have an increased risk of giving birth to low birth weight infants (Han et al., 2011) and the low prevalence of underweight mothers in our study population, women with a pre-pregnancy BMI of < 18.5 kg/m2 (n = 28) were excluded from this analysis.

2.5. Other measures

Women self-reported demographics, lifestyle characteristics, and medical history via prenatal and postpartum questionnaires. Trained study staff also conducted a medical record review to collect maternal weight throughout pregnancy. Gestational weight gain was calculated as the difference between the final reported weight from pregnancy and the self-reported pre-pregnancy weight. Gestational weight gain was then categorized as inadequate, adequate, or excessive according to Institute of Medicine (IOM) guidelines (Rasmussen and Yatkine, 2009).

Duration of breastfeeding was obtained via repeated telephone questionnaires, conducted when the infant was 4, 8, and 12 months of age, in which women were asked whether they were still breastfeeding and, if they were not, the date or the age of the infant when they had stopped breastfeeding. Additionally, if they were still breastfeeding, they were subsequently asked if they had ever fed their baby anything other than breast milk to establish whether a child had been exclusively breast fed up until that point. For statistical analyses, a breastfeeding variable was derived as the total number of months an infant was breastfed. In regression analyses, this variable was set to the lesser of the following two values: the total number of months the infant was breast fed and the age of the infant when the anthropometry values were collected.

2.6. Statistical analyses

All analyses were completed using R: A Language and Environment for Statistical Computing, version 3.4.0 (R Foundation for Statistical Computing, Vienna, Austria). A threshold of p < 0.05 was used to define main effects as statistically significant. To improve sensitivity to identify potential effect modification, a threshold of p < 0.10 was used to identify interactions for further consideration. We assessed distributions and univariate statistics of baseline characteristics, maternal urinary arsenic concentrations, and anthropometric measures at various ages.

After initial univariate analysis, we used natural log-transformed tAs as the exposure for all analyses because tAs was right-skewed. We then assessed the linearity of each dose-response relationship between tAs and each outcome through visual inspection of scatter plots of the data. We did not find evidence of any non-linear dose-response relationships between ln-transformed maternal urinary tAs (Fig. S1) and each outcome.

We used a series of linear regression models to fit each anthropometric outcome at each time period on ln-transformed tAs. Models were fit in the overall sample and then stratified by maternal pre-pregnancy weight status (normal weight vs. overweight and obese) or infant sex. In all models, variables associated with maternal urinary tAs or any anthropometric outcome at the p < 0.10 level in bivariate analyses were included as covariates to increase sensitivity to identify potential confounders of the association. The adjusted non-stratified linear regression models included maternal pre-pregnancy BMI (continuous), gestational weight gain (categorical; insufficient, sufficient, or excessive), maternal education (ordinal, with the four categories listed in Table 1 assigned values of 1–4, respectively), maternal smoking status (binary; ever smoked or never smoked), maternal parity (ordinal with parity of 3 or more set to 3), infant sex (binary), infant gestational age at delivery (continuous), and months breastfed (continuous). Additional variables that were considered but that did not achieve the stated threshold for inclusion in subsequent models included maternal age (continuous) and delivery mode (binary; vaginal delivery or cesarean section). Models stratified by maternal pre-pregnancy weight status did not adjust for maternal pre-pregnancy BMI. Models stratified by infant sex did not adjust for infant sex. Missing values for maternal pre-pregnancy BMI (n = 30) were assigned the study population mean (26.11 kg/m2). Missing values for months breast fed (n = 123) were assigned the study population median (7.98 months). Missing values for maternal education (n = 66), maternal smoking status (n = 69), and gestational weight gain category (n = 36) were treated with a missing variable indicator. These methods assume that the missingness of a given covariate is not related to the exposure of interest, the violation of which could lead to biased estimates (Groenwold et al., 2012). This assumption is valid in these analyses as the missingness of these values is not associated with arsenic exposure (data not shown).

Table 1.

Selected subject characteristics.

Variable All Participants (n = 760) Missing Values (n)
Pre-pregnancy BMI (kg/m2) [mean (sd)] 26.11 (5.52) 30
Maternal pre-pregnancy weight status [n (%)] 30
 Normal (18.5 ≤ BMI < 25) 388 (53.2)
 Overweight (25 ≤ BMI < 30) 200 (27.4)
 Obese (BMI ≥ 30) 142 (19.5)
 Missing 30 (3.9)
 Maternal age (yrs) [mean (sd)] 31.09 (4.84) 0
Maternal education level [n (%)] 66
 High school graduate or lower 78 (11.2)
 Some college 135 (19.5)
 College graduate 289 (41.6)
 Postgraduate schooling 192 (27.7)
 Missing 66 (8.7)
Previous pregnancies [n (%)] 0
 0 301 (39.6)
 1 290 (38.2)
 2 107 (14.1)
 ≥3 62 (8.2)
Maternal Smoking Status [n (%)] 69
 Never smoker 591 (85.5)
 Ever smoker 100 (14.5)
 Missing 69 (9.1)
Gestational weight gain (lbs) [mean (sd)] 35.15 (14.88) 33
IOM gestational weight gain category [n (%)] 63
 Adequate 196 (27.1)
 Inadequate 79 (10.9)
 Excessive 449 (62.0)
 Missing 36 (4.7)
Infant delivery mode [n (%)] 0
 Vaginal 515 (67.8)
 Cesarean section 245 (32.2)
Infant Sex [n (%)] 0
 Female 379 (49.9)
 Male 381 (50.1)
Birth Outcomes [mean (sd)]
 Gestational age (wks) 39.55 (1.39) 0
 Length WHO Z score 0.78 (1.31) 0
 Weight WHO Z score 0.35 (0.97) 0
 Weight-for-Length (WFL) WHO Z score −0.38 (1.42) 0
 Head circumference WHO Z score 0.41 (1.18) 4
Duration of breast feeding (mo) (median [IQR]) 7.98 [3.00, 12.16] 123
Open in a new tab

Models with multiplicative interaction terms between the ln-transformed tAs levels and either maternal pre-pregnancy weight status or infant sex were used to formally assess effect modification by those variables; a likelihood ratio test was used to determine statistical significance of the multiplicative terms. As an exploratory analysis, for anthropometry outcomes for which the likelihood ratio test was suggestive (p < 0.10) of effect modification by either maternal weight status or infant sex, we estimated the associations with tAs exposure stratified by both maternal weight status and infant sex. To test for the significance of a 3-way interaction between maternal weight status, infant sex, and tAs exposure, we fit a multivariable regression with main effect terms, 2-way multiplicative interaction terms, and a 3-way multiplicative interaction term, adjusted for covariates; we assessed the statistical significance of the 3-way multiplicative interaction by comparing the model with a 3-way multiplicative interaction term to a nested model without a 3-way multiplicative interaction term using a likelihood ratio test.

In addition, we fit the longitudinal age- and sex-adjusted Z-score data for weight, length, weight-for-length, and head circumference collected at birth, 2 weeks, 1 month, 2 months, 4 months, 6 months, 9 months, and one year in linear mixed effects models adjusted for the same covariates described above. Models were fit using the lme4 package (version 1.1–13) in R (Bates et al., 2015) and included fixed effects for ln-transformed tAs levels and covariates as well as a random intercept term to account for subject-specific anthropometric variability.

Change in growth trajectory was assessed by fitting piecewise linear mixed effects models for weight, length, weight-for-length, and head circumference measures collected at 2 weeks, 1 month, 2 months, 4 months, 6 months, 9 months, and one year adjusted for the same covariates described above. Raw anthropometry measures were used due to the non-linear association between WHO Z scores and age over the first year in our study population (Fig. S2). Models included fixed effects for anthropometry at birth (continuous), ln-transformed tAs (continuous), and covariates as well as random intercept and slope terms to account for subject-specific anthropometric variability. Interaction terms between infant sex and infant age as well as between breast feeding duration and infant age were included to allow for varying growth trajectories based on sex and breastfeeding status. Piecewise models were used due to a visually observed change in slope in the first year of life (Fig. S3). Change points were selected based on the age (assessed at ½ month intervals) that minimized the residual sums of squares (Bai, 1997) for unadjusted models, however selected change points did not differ when adjusting for all covariates used for the longitudinal model. Unadjusted piecewise models stratified by infant sex were used to obtain the mean growth trajectories for both segments of the model for both sexes in the study population.

Anthropometric outcomes that were identified as having a significant association with ln-transformed tAs levels were also tested for an association with maternal urinary inorganic arsenic (iAs), MMA, DMA, and tAs on a categorical scale with 3 levels. Due to the large number of subjects with iAs and MMA below the LOD, these exposures were categorized with all subjects below the LOD being assigned to the lowest level of exposure (35.5% and 23.3%, respectively), all subjects below the median of exposure for subjects above the LOD being assigned the second level of exposure, and all subjects at or above the median exposure for subjects above the LOD being assigned the highest level of exposure. For DMA and tAs, levels of exposure were established based on exposure < 33rd percentile, ≥33rd percentile and < 67th percentile, and ≥67th percentile. Models with arsenic exposure treated as three ordered quantiles (approximating tertiles) used the same covariates described above for models with arsenic exposure treated as continuous ln-transformed tAs.

Several sensitivity analyses were completed to assess the robustness of the results. Given that the metabolism of arsenolipids and arsenosugars from seafood consumption can contribute to measured levels of DMA and, subsequently, tAs (Francesconi et al., 2002; Raml et al., 2005; Taylor et al., 2017), multivariable regression models assessing associations of maternal urinary tAs with infant anthropometry were repeated 1) restricting to women who did not report any seafood intake within the two days prior to the urine sample collection (n = 653) and 2) adjusting for log-transformed maternal second trimester urinary arsenobetaine concentration (n = 760) to assess possible confounding of observed results by maternal seafood consumption. Multivariable regression models assessing associations of maternal urinary tAs with infant anthropometry were also repeated: 1) restricting to women who did not report any rice intake within the two days prior to the urine sample collection (n = 525), and 2) adjusting for creatinine for the subset of women with urinary creatinine concentrations available (n = 646), 3) restricting to infants who were exclusively breastfed at 6 weeks (n = 270), 4) restricting to infants who were not large for gestational age (at or above the 90th percentile per Fenton and Kim 2013 growth curves (Fenton and Kim, 2013); n = 691), 5) restricting to infants who were born at a minimum of 37 weeks gestation and who were not small for gestational age (at or below the 10th percentile (Fenton and Kim, 2013); n = 693), 6) adjusting for whether or not rice cereal (a post-natal source of arsenic exposure) had been introduced at the time of the well child visit when anthropometry data were collected for the subset of infants with completed questionnaire data on the introduction of rice cereal (n = 342), and 7) restricting to women who did not have gestational diabetes during this pregnancy, defined as a 1-hr glucose challenge test (GCT) result ≥ 200 mg/dL or failing the 1-hr GCT (140–199 mg/dL) with two or more high values on the oral glucose tolerance test (OGTT) based on the American Diabetes Association criteria for a normal OGTT of blood glucose ≤ 95 mg/dL at baseline, ≤ 180 mg/dL at 1 h, ≤ 155 mg/dL at 2 h, and ≤140 mg/dL at 3 h (n = 703). Analyses testing the association between maternal urinary tAs and attained anthropometry were conducted with and without adjustment for anthropometry at birth in order to help elucidate potential mediation of associations by birth anthropometry. In addition, adjusting for baseline measurements when assessing change in measurements over time can introduce bias when the exposure and baseline measurement are associated (Glymour et al., 2005), so we also compared growth trajectory analysis results with and without adjusting for anthropometry at birth.

3. Results

In total, 760 mother-infant pairs had both second trimester urinary arsenic and anthropometry measurements at birth and during some or all of the first year of life available. Table 1 presents summary statistics for subject background characteristics and birth outcomes for the study population. Over 45% of mothers were classified as overweight or obese based on their pre-pregnancy BMI. The median second trimester urinary total arsenic (tAs) concentration for the 760 mothers included in the analysis was 3.96 μg/L with DMA accounting for greater than 80% of urinary tAs in more than 50% of the samples (Table 2). When comparing women who were included in this study to the entire population-based cohort, women in this study had statistically significantly higher gestational weight gain and infants had statistically significantly higher gestational age, length WHO Z score, and weight WHO Z score at birth (Table S1).

Table 2.

Distribution of arsenic exposure for 760 mother-infant pairs.

Arsenic Measure Concentration (μg/L)
Median [IQR]		 Proportion of tAs
Median [IQR]		 Below LODa
n (%)
iAsb 0.36 [0.16, 0.81] 0.10 [0.06, 0.17] 270 (35.5)
MMA 0.34 [0.13, 0.66] 0.08 [0.05, 0.12] 177 (23.3)
DMA 2.84 [1.39, 5.07] 0.81 [0.70, 0.86] 3 (0.4)
tAsc 3.96 [2.02, 6.72] 1.00 [1.00, 1.00] 2 (0.3)
Open in a new tab
a

Number of samples below the limit of detection (LOD).

b

iAs reflects the sum of AsIII and AsV measured in each sample where the number of samples below the LOD reflects the number of samples in which both AsIII and AsV were below the LOD.

c

tAs reflects the sum of iAs (AsIII + AsV), MMA, and DMA measured in each sample where the number of samples below the LOD reflects the number of samples in which AsIII, AsV, MMA, and DMA were all below the LOD.

In the overall study population, natural log-transformed maternal second trimester urinary tAs concentration was positively associated with infant length such that each doubling of urinary tAs was associated with a 0.05 (95% CI: 0, 0.09; p = 0.044) higher length WHO Z score over the first year of life (Table 3). In analyses assessing the association between tAs and attained length at each time point, each doubling of urinary tAs was associated with a 0.07 (95% CI: 0.01, 0.13; p = 0.025) higher length WHO Z score at 2 weeks but no statistically significant association was observed at subsequent timepoints.

Table 3.

Adjusted parameter estimates (95% confidence interval) for the change in anthropometry WHO Z scores with each doubling of total maternal second trimester urinary arsenic for all participants.

Outcome Timepoint Overalla
n Beta (95% CI) p
Weight 2 weeks 686 0.02 (−0.02, 0.06) 0.36
2 months 744 0.02 (−0.03, 0.07) 0.45
4 months 722 0.01 (−0.04, 0.06) 0.63
6 months 720 0.01 (−0.04, 0.06) 0.70
9 months 691 0.02 (−0.03, 0.07) 0.51
12 months 691 0.01 (−0.04, 0.06) 0.67
Repeated Measuresb 760 (5,376) 0.01 (−0.03, 0.05) 0.52
Length 2 weeks 585 0.07 (0.01, 0.13) 0.025
2 months 741 0.04 (−0.01, 0.10) 0.11
4 months 722 0.02 (−0.04, 0.07) 0.50
6 months 717 0.02 (−0.03, 0.08) 0.40
9 months 688 0.04 (−0.02, 0.09) 0.24
12 months 688 0.05 (−0.01, 0.11) 0.10
Repeated Measuresb 760 (5,238) 0.05 (0.00, 0.09) 0.044
WFL 2 weeks 584 −0.04 (−0.11, 0.02) 0.20
2 months 738 −0.03 (−0.10, 0.03) 0.32
4 months 722 0.00 (−0.06, 0.06) 1.00
6 months 717 0.00 (−0.06, 0.06) 0.90
9 months 688 0.00 (−0.06, 0.05) 0.97
12 months 688 −0.02 (−0.07, 0.03) 0.51
Repeated Measuresb 760 (5,234) −0.03 (−0.07, 0.02) 0.22
HC 2 weeks 577 −0.03 (−0.09, 0.03) 0.29
2 months 727 0.01 (−0.04, 0.06) 0.80
4 months 704 0.00 (−0.05, 0.05) 0.90
6 months 697 −0.01 (−0.06, 0.04) 0.69
9 months 667 0.01 (−0.05, 0.06) 0.83
12 months 663 −0.01 (−0.07, 0.04) 0.62
Repeated Measuresb 756 (5,109) −0.01 (−0.05, 0.04) 0.80
Open in a new tab
a

Results are from linear regression models fitting weight, length, weight-for-length (WFL), and head circumference (HC) WHO Z scores (WHO Multicentre Growth Reference Study Group, 2006) on log transformed maternal second trimester urinary tAs. Models are adjusted for gestational age (continuous), infant sex (binary), maternal pre-pregnancy BMI (continuous), IOM gestational weight gain category (Rasmussen and Yatkine, 2009) (categorical), maternal education level (ordinal with 4 levels listed in Table 1), parity (ordinal with parity ≥ 3 grouped together), duration of breast feeding up to the time of anthropometry measurement (continuous), and maternal smoking status (binary; ever or never).

b

Results are from a linear mixed effects model fitting a random intercept for each infant. Sample size reflects the total number of infants included in the analysis and, in parentheses, the total number of measurements included in the model. Each infant contributed between 2 and 8 measurements (collected at birth, 2 weeks, 1 month, 2 months, 4 months, 6 months, 9 months, and/or 12 months).

In analyses that assessed in utero arsenic exposure by three ordered quantiles of specific arsenic species, individuals in DMA quantile 3 had, on average, a 0.24 increase (95% CI: 0.09, 0.39; ptrend = 0.0019) in attained length WHO Z score compared to those in quantile 1 when examining the relationship in repeated measures analysis as well as at several timepoints during the first year of life (Table S2). There was also a statistically significant trend for the positive association with iAs quantile with length Z score at 2 weeks of age (ptrend = 0.045), but not thereafter. There were no significant associations between MMA quantiles and attained length.

There was no statistically significant interaction identified between urinary tAs and maternal pre-pregnancy weight status (p = 0.35; Table S3), nor between urinary tAs and infant sex (p = 0.77; Table 4) in relation to infant length over the first year of life.

Table 4.

Adjusted parameter estimates (95% confidence interval) for the change in anthropometry WHO Z scores with each doubling of total maternal second trimester urinary arsenic stratified by infant sex.

Outcome Timepoint Femalea
 Malea
 pintb
n Beta (95% CI) p n Beta (95% CI) p
Weight 2 weeks 342 0.02 (−0.04, 0.08) 0.55 344 0.00 (−0.06, 0.07) 0.88 0.70
2 months 369 0.02 (−0.04, 0.08) 0.54 375 0.02 (−0.06, 0.10) 0.60 0.98
4 months 356 0.04 (−0.02, 0.10) 0.22 366 −0.01 (−0.10, 0.07) 0.72 0.25
6 months 357 0.05 (−0.02, 0.11) 0.16 363 −0.03 (−0.11, 0.05) 0.43 0.10
9 months 343 0.05 (−0.02, 0.11) 0.15 348 −0.02 (−0.09, 0.06) 0.65 0.16
12 months 344 0.03 (−0.03, 0.09) 0.31 347 −0.01 (−0.09, 0.06) 0.73 0.32
Repeated Measuresc 379 (2,670) 0.03 (−0.02, 0.08) 0.23 381 (2,706) −0.01 (−0.07, 0.05) 0.82 0.31
Length 2 weeks 289 0.04 (−0.05, 0.12) 0.38 296 0.10 (0.01, 0.18) 0.022 0.19
2 months 367 0.05 (−0.03, 0.13) 0.20 374 0.05 (−0.04, 0.13) 0.27 0.96
4 months 356 0.01 (−0.07, 0.09) 0.78 366 0.02 (−0.06, 0.10) 0.55 0.74
6 months 355 0.02 (−0.05, 0.10) 0.57 362 0.03 (−0.06, 0.11) 0.53 0.91
9 months 340 0.04 (−0.05, 0.12) 0.39 348 0.04 (−0.05, 0.12) 0.41 0.95
12 months 341 0.07 (−0.01, 0.15) 0.10 347 0.04 (−0.05, 0.13) 0.41 0.65
Repeated Measuresc 379 (2,594) 0.05 (−0.02, 0.11) 0.15 381 (2,644) 0.05 (−0.01, 0.12) 0.12 0.77
WFL 2 weeks 289 −0.04 (−0.13, 0.05) 0.42 295 −0.09 (−0.19, 0.01) 0.068 0.27
2 months 367 −0.03 (−0.12, 0.05) 0.42 371 −0.04 (−0.13, 0.06) 0.45 0.82
4 months 356 0.04 (−0.03, 0.12) 0.29 366 −0.04 (−0.13, 0.05) 0.36 0.12
6 months 355 0.04 (−0.03, 0.12) 0.28 362 −0.06 (−0.15, 0.03) 0.19 0.056
9 months 340 0.04 (−0.03, 0.12) 0.27 348 −0.05 (−0.13, 0.03) 0.23 0.069
12 months 341 0.00 (−0.07, 0.07) 0.98 347 −0.04 (−0.12, 0.04) 0.31 0.36
Repeated Measuresc 379 (2,594) 0.00 (−0.05, 0.06) 0.91 381 (2,640) −0.07 (−0.13, 0.00) 0.041 0.063
HC 2 weeks 285 −0.07 (−0.15, 0.01) 0.087 292 −0.02 (−0.10, 0.07) 0.69 0.45
2 months 358 −0.01 (−0.08, 0.05) 0.66 369 0.03 (−0.04, 0.10) 0.42 0.46
4 months 345 −0.02 (−0.09, 0.05) 0.54 359 0.02 (−0.05, 0.10) 0.58 0.56
6 months 343 0.01 (−0.06, 0.08) 0.83 354 −0.02 (−0.10, 0.05) 0.55 0.42
9 months 331 0.03 (−0.05, 0.10) 0.51 336 0.00 (−0.08, 0.07) 0.92 0.45
12 months 330 0.02 (−0.06, 0.10) 0.59 333 −0.04 (−0.12, 0.04) 0.28 0.14
Repeated Measuresc 376 (2,527) 0.00 (−0.06, 0.06) 0.99 380 (2,582) 0.00 (−0.07, 0.06) 0.89 0.81
Open in a new tab
a

Results are from linear regression models fitting weight, length, weight-for-length (WFL), and head circumference (HC) WHO Z scores (WHO Multicentre Growth Reference Study Group, 2006) on log transformed maternal second trimester urinary tAs and stratified by infant sex. Models are adjusted for gestational age (continuous), maternal pre-pregnancy BMI (continuous), IOM gestational weight gain category (Rasmussen and Yatkine, 2009) (categorical), maternal education level (ordinal with 4 levels listed in Table 1), parity (ordinal with parity ≥ 3 grouped together), duration of breast feeding up to the time of anthropometry measurement (continuous), and maternal smoking status (binary; ever or never).

b

P values reflect the p value from a likelihood ratio test comparing a model with an interaction between maternal arsenic and infant sex to a nested model with no interaction term.

c

Results are from a linear mixed effects model fitting a random intercept for each infant. Sample size reflects the total number of infants included in the analysis and, in parentheses, the total number of measurements included in the model. Each infant contributed between 2 and 8 measurements (collected at birth, 2 weeks, 1 month, 2 months, 4 months, 6 months, 9 months, and/or 12 months).

Maternal second trimester urinary tAs concentration was not related to infant attained weight, weight-for-length, or head circumference during the first year of life in the full sample (Table 3). No evidence was found of an interaction between urinary tAs and infant sex with respect to attained weight or head circumference over the first year of life. However, in analyses stratified by infant sex, urinary tAs was negatively associated with weight-for-length Z score over the first year of life exclusively in males, but not females (Table 4). In males, every doubling of urinary tAs was associated with a 0.07 decrease (95% CI: −0.13, 0; p = 0.041) in weight-for-length Z score over the first year of life (pinteraction = 0.063); among females this was 0 (95% CI: −0.05, 0.06; p = 0.91).

When assessing in utero arsenic exposure by arsenic species, the negative association between maternal urinary arsenic and male attained weight-for-length Z score was strongest when assessing arsenic exposure as MMA quantiles with quantile 3 being associated with a 0.22 decrease (95% CI: −0.45, 0; ptrend = 0.039) in weight-for-length Z score relative to quantile 1 over the first year of life (Table S4). Weak to null associations were observed between male attained weight-for-length Z score and iAs (ptrend = 0.36) or DMA (ptrend = 0.15) quantiles over the first year of life.

Further analyses failed to identify evidence of a three-way multiplicative interaction between urinary tAs, infant sex, and maternal pre-pregnancy weight status with respect to weight-for-length Z score over the first year of life (pinteraction = 0.23; Table S5). No evidence was found of an interaction between urinary tAs and maternal pre-pregnancy weight status with respect to attained weight, weight-for-length, or head circumference over the first year of life (Table S3).

In the study population, the average growth rate over the first 3.5 months of life was 3.50 cm per month in females and 3.76 cm per month in males (Table S6). In adjusted piecewise linear mixed effects models assessing the association between maternal urinary second trimester tAs concentration and infant growth trajectory, urinary tAs was negatively associated with change in length for the first 3.5 months of life (Table 5). Overall, each doubling of maternal urinary tAs was associated with a 0.07 cm per month lower length growth (95% CI: −0.12, −0.02; p = 0.0051) over the first 3.5 months of life. No association was observed between change in length and urinary tAs after 3.5 months (p = 0.26). There was no evidence of an interaction between urinary tAs and infant sex (pinteraction = 0.18) or maternal pre- pregnancy weight status (pinteraction = 0.73; Table S7) in relation to change in length over the first 3.5 months of life.

Table 5.

Adjusted parameter estimates (95% confidence interval) for the change in growth trajectory associated with each doubling of total maternal second trimester urinary arsenic, overall and by infant sex.

Outcome Age (months) Overall (n = 760) a
 Female (n = 379) b
 Male (n = 381) b
 Pint c
Beta (95% CI) p Beta (95% CI) p Beta (95% CI) p
Weight (g/mo) 0–3.5 2.52 (−9.68, 14.71) 0.69 19.27 (3.12, 35.41) 0.019 −14.04 (−32.21, 4.13) 0.13 0.035
3.5–12 1.06 (−6.15, 8.28) 0.77 3.58 (−5.74, 12.90) 0.45 −2.04 (−13.02, 8.95) 0.72 0.92
Length (cm/mo) 0–3.5 −0.07 (−0.12, −0.02) 0.0051 −0.03 (−0.10, 0.03) 0.30 −0.11 (−0.18, −0.03) 0.0039 0.18
3.5–12 0.01 (−0.01, 0.03) 0.26 0.02 (−0.01, 0.05) 0.16 0.00 (−0.03, 0.04) 0.79 0.42
WFL (g/cm/mo) 0–3 0.17 (−0.06, 0.40) 0.15 0.33 (0.02, 0.64) 0.036 0.04 (−0.30, 0.37) 0.83 0.20
3–12 −0.01 (−0.10, 0.08) 0.84 0.01 (−0.11, 0.13) 0.90 −0.04 (−0.17, 0.10) 0.61 0.42
HC (cm/mo) 0–3.5 0.01 (−0.02, 0.04) 0.56 0.01 (−0.03, 0.05) 0.63 0.01 (−0.03, 0.04) 0.64 0.41
3.5–12 0.00 (−0.01, 0.01) 0.69 0.01 (−0.01, 0.02) 0.29 −0.01 (−0.03, 0.00) 0.062 0.052
Open in a new tab
a

Results are from piecewise linear mixed effects models fitting each anthropometry measure on log transformed maternal second trimester urinary tAs. Parameters reflect the association between tAs and anthropometry growth trajectory, as longitudinally modeled by the multiplicative term between infant age and tAs exposure. Models are adjusted for anthropometry measurement at birth (continuous), gestational age (continuous), infant sex (with an additional interaction term between infant sex and infant age), maternal pre-pregnancy BMI (continuous), IOM gestational weight gain category (categorical), maternal education level (ordinal with 4 levels listed in Table 1), parity (ordinal with parity ≥ 3 grouped together), duration of breast feeding up to the time of anthropometry measurement (continuous; with an additional interaction term between duration of breast feeding and infant age), and maternal smoking status. The changepoint for the piecewise model was selected based on the age (assessed at ½ month intervals) that minimized the residual sums of squares for unadjusted models.

b

Results are from piecewise linear mixed effects models fitting each anthropometry measure on log transformed maternal second trimester urinary tAs and stratified by infant sex. Models are adjusted for all variables described in footnote a with the exception of infant sex.

c

P value from a likelihood ratio test comparing a model with all the covariates and interaction terms described in footnote a as well as a 3-way interaction term between maternal arsenic, infant age, and infant sex and a 2-way interaction term between maternal arsenic and infant sex to a nested model with no 3-way interaction term and no 2-way interaction term between maternal arsenic and infant sex.

When comparing infant length growth during the first 3.5 months of life for those in the 3rd vs. 1st quantile of in utero arsenic exposure, iAs and MMA were both associated with decreased growth (Table S8); for iAs, the decrease was 0.12 cm per month (95% CI: −0.22, −0.02; ptrend = 0.013) and for MMA it was 0.13 cm per month (95% CI: −0.24, −0.02; ptrend = 0.0097). This negative association with infant length growth was not observed when assessing exposure by DMA quantiles (ptrend = 0.10).

Over the first 3.5 months of life, females in this study gained, on average, 848.23 g per month and males gained, on average, 1013.23 g per month (Table S6). There was evidence of an interaction between urinary tAs and infant sex with respect to change in weight over the first 3.5 months of life (Table 5; pinteraction = 0.035). Adjusted longitudinal analyses stratified by infant sex demonstrated that urinary tAs was positively associated with change in weight in females, but not in males (p = 0.13). Among females, every doubling of urinary tAs concentration was associated with a 19.27 (95% CI: 3.12, 35.41; p = 0.019) g per month increase in weight gain over the first 3.5 months of life. There was no evidence of an association between weight gain and tAs after 3.5 months among either females (p = 0.45) or males (p = 0.72).

When assessing weight gain during the first 3.5 months of life in relation to in utero arsenic exposure quantiles in females, being in the 3rd vs 1st quantile of MMA was associated with a 53.03 g per month increase in weight gain (95% CI: 12.05, 94.01; ptrend = 0.0091) and being in the 3rd vs. 1st quantile of DMA was associated with a 43.20 g per month increase in weight gain (95% CI: 5.88, 80.53; ptrend = 0.02) (Table S9). This positive association with infant weight gain was not observed when assessing exposure by iAs quantiles (ptrend = 0.32).

All findings were consistent in analyses restricted to women who had not eaten seafood in the two days prior to urine collection (n = 653; Tables S10–11), adjusting for log-transformed maternal second trimester urinary arsenobetaine concentration (n = 760; Tables S12–13), restricted to women who had not eaten rice in the two days prior to urine collection (n = 525; Tables S14–15), restricted to women who did not have gestational diabetes during this pregnancy (n = 703; Tables S16–17), restricted to infants who were not large for gestational age (n = 691; Tables S18–19), restricted to infants who were born after at least 37 weeks gestation and were not small for gestational age (n = 693; Tables S20–21), and restricted to infants who were exclusively breast feed at 6 weeks (n = 270; Tables S22–23). Findings were also consistent when adjusting for maternal urinary creatinine concentrations (n = 646; Tables S24–25) and adjusting for whether rice cereal had been introduced at the time anthropometry measurements were collected (n = 342; Tables S26–27). Attained anthropometry analyses were consistent but attenuated in analyses adjusting for anthropometry at birth (Table S28). Growth trajectory findings were consistent in analyses that did not adjust for anthropometry at birth (Table S29).

4. Discussion

Within this sample of 760 mother-infant pairs in New Hampshire, we found modest associations between in utero arsenic exposure, as measured by maternal second trimester urinary tAs concentrations, and infant anthropometry WHO Z scores and growth trajectory over the first year of life. An overall modest positive association was observed between maternal pregnancy urine tAs levels and infant length, as assessed by WHO Z scores, throughout the entire first year of life. A 0.05 increase in Z score corresponds to an approximately 0.12 cm increase in males and 0.13 cm increase in females at 12 months. This association was attenuated after adjusting for length WHO Z score at birth, suggesting that the observed association may be mediated by anthropometry at birth. As this association only achieved statistical significance at 2 weeks in timepoint specific models, it is possible that the association observed in the repeated measures analysis is primarily driven by this single timepoint.

Higher maternal urinary tAs levels were also negatively associated with change in length over the first 3.5 months of life, but not thereafter. This negative association with length growth trajectory may represent a true effect of arsenic exposure on early growth or, given previous findings associating biomarkers of in utero arsenic exposure with longer birth length in this population (Gilbert-Diamond et al., 2016), this may reflect “regression to the mean”, the statistical tendency for values that are more extreme to be followed by values that are more moderate (Barnett et al., 2004). It is important to note that given the large number of statistical tests being conducted and the use of a nominal threshold of statistical significance of p < 0.05, it is possible that identified associations may be spurious and due to chance.

Interestingly, observed associations between in utero arsenic exposure and infant length were not found to be exclusively driven by DMA, the arsenic metabolite that accounted for over 80% of urinary tAs in more than 50% of participants in this study. This suggests that exposure to iAs and MMA may be more important for some anthropometric outcomes than total arsenic exposure (iAs and its metabolites). Further work is needed to better understand the association between in utero exposure to specific arsenic species and subsequent anthropometric outcomes.

In assessing possible effect modification by infant sex, we identified a persistent negative association between maternal urinary total arsenic levels and weight-for-length over the entire first year of life among male infants only. This decreased adiposity may be due to arsenic-associated greater length. Interestingly, in females, there was a positive association between urine total arsenic levels and the rate of weight gain over the first 3.5 months of life, but this did not translate into any statistically significant differences in weight or weight-for-length WHO Z scores over the first year of life. Previous studies have observed sex-specific associations between arsenic and growth (Gilbert-Diamond et al., 2016; Saha et al., 2012). One possible explanation for these sex-specific associations is mediation by the placental expression of AQP9, a gene encoding a transporter involved in cellular uptake of arsenic that has been reported to vary with arsenic exposure in a sex-specific manner (Winterbottom et al., 2017).

Growth during infancy has been associated with risk of metabolic disease during adulthood. Therefore, altered infant growth related to in utero arsenic exposure may have long-term health consequences. Persistently lowered BMI throughout infancy (< 15 kg/m2) has been associated with an increased risk of impaired glucose tolerance in early adulthood (Bhargava et al., 2004), and when that lowered BMI is followed by a rapid increase in BMI during childhood it has been associated with type 2 diabetes in later adulthood (Eriksson et al., 2015). Decreased change in length in the first 3 months of infancy has also been associated with an increased risk of type 2 diabetes in adulthood in a dose dependent fashion among infants with birth weights above 3.5 kg, particularly when followed by a rapid increase in BMI after 2 years of age (Eriksson et al., 2003). Given that our study observed very modest changes in infant length (a 0.07 cm per month decrease) relative to those studied in Eriksson et al. where differences in growth rate were assessed in 1 cm per month increments, our observed deviations in growth trajectory may not be substantially associated with metabolic outcomes in adulthood.

Saha et al. (2012) identified a weak negative association between in utero arsenic exposure and infant weight and length over the first year of life. These disparate findings may be due to differences in study population characteristics. For example, our studies differed greatly in terms of maternal BMI (average of 20.1 kg/m2 in the Saha study compared to 26.3 kg/m2 in our study population). In addition, the median urine tAs concentration was much higher in the Bangladeshi study population examined by Saha et al. (84.0 μg/L compared to 3.83 μg/L in our study population). The notion that arsenic exposure may have a differential relationship with growth dependent on dose is supported by previous in vitro research on cultured osteoblasts which are critical to bone formation. Research by Xu et al. reported that, in an in vitro model, lower levels of arsenic trioxide (ATO) exposure (0.25, 0.5, and 1 μM) enhance osteoblast viability in culture while higher levels of ATO exposure (5, 10, and 20 μM) reduce osteoblast viability in culture (Xu et al., 2014). However, it is currently unknown how arsenic trioxide exposure relates to tAs exposure or how this threshold effect might translate to a threshold of total arsenic concentration in urine. As the lower levels of total urinary arsenic concentrations seen in our study population encompass those commonly seen throughout the United States, our findings may be more relevant to the United States population than those from heavily exposed Bangladeshi populations.

Other studies relating high dose prenatal arsenic exposure to postnatal growth at 5 and 7 years of age (Agay-Shay et al., 2015; Gardner et al., 2013) did not identify an association between in utero arsenic exposure and child anthropometry. It will therefore be important to assess the association between prenatal arsenic exposure and later childhood growth in our population with relatively lower doses of in utero arsenic exposure.

Our study strengths include having speciated measures of arsenic exposure: urinary inorganic arsenic and its metabolites which act as sensitive biomarkers of toxicologically relevant in utero arsenic exposure. Our study was also strengthened by having access to clinical anthropometry measurements collected at multiple time points across the first year of infancy. However, clinical measures of infant length may be prone to error due to variability in measurement techniques across health care providers (Wood et al., 2013). In addition, instruments (for example, scales) from multiple clinical practices are not standardized nor uniformly calibrated as would be the case in a research setting. Given that our study relied on clinical measurements obtained from medical record review of infant well child visits, added variability around anthropometric measurements may have diminished our power to detect an association between biomarkers of in utero arsenic exposure and infant growth.

Our study was limited by arsenic exposure only being measured at one time during pregnancy; however one study conducted in Native American populations with relatively low levels of arsenic exposure suggests that measures of urinary arsenic are relatively stable over time (Navas-Acien et al., 2009). Our study was also limited by not assessing biomarkers of postnatal arsenic exposure in infants, however our findings were robust in sensitivity analyses restricted to infants who were exclusively breast fed, a subgroup without demonstrable postnatal arsenic exposure (Carignan et al., 2015, 2016; Fängström et al., 2008). As women have been shown to underestimate their pre-pregnancy weight via self-report, causing some overweight mothers to be misclassified as normal weight prior to pregnancy (Han et al., 2016), our reliance on self-reported maternal pre-pregnancy weight may have led to some subjects being misclassified with regards to pre-pregnancy weight status. Such misclassification may have obscured potential differences between women of normal as compared to overweight and obese pre-pregnancy weight statuses, thus limiting our ability to test for modification by pre-pregnancy weight status. Our study was also limited by sample size, particularly in analyses stratified by infant sex and/ or maternal pre-pregnancy weight status. Given the limited number of infants in each group following stratification, we may have been underpowered to detect effect modification by these factors.

5. Conclusions

Our findings suggest that, on average, infants exposed to higher in utero arsenic attained slightly longer length through the first 12 months of life, despite having slower linear growth in the first 3.5 months after birth. This suggests that the previously reported positive association between in utero arsenic exposure and birth length among study infants persists through the first year of life. We did not observe any other statistically significant associations between in utero arsenic exposure and child anthropometry across the full study population. Future studies in our cohort are needed to better understand how both prenatal and postnatal arsenic exposure may continue to affect anthropometry and growth trajectories throughout infancy and into childhood.

Supplementary Material

supplementMtrl

Acknowledgments

We would like to thank the study participants, staff, and collaborators on the New Hampshire Cohort Study. The study was funded in part: National Institutes of Health (NIH) National Institute of Environmental Health Sciences (United States) grants P42ES007373, P01ES022832, and P42ES004940, NIH National Institute of General Medical Sciences (United States) grant P20GM104416, Environmental Protection Agency (United States) grant RD83544201, Burroughs-Wellcome/Dartmouth Big Data in the Life Sciences Training Program. No funding bodies had any role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. This manuscript was finalized while KLC was serving as a program director at the National Science Foundation.

Abbreviations

As

arsenic

tAs

total arsenic

iAs

inorganic arsenic

MMA

monomethylarsonic acid

DMA

dimethylarsinic acid

WHO

World Health Organization

BMI

body mass index

LOD

limit of detection

IOM

Institute of Medicine

Footnotes

Competing financial interests declaration

The authors declare they have no actual or potential competing financial interests.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.envres.2019.108604.

References

  1. Agay-Shay K, Martinez D, Valvi D, Garcia-Esteban R, Basagaña X, Robinson O, et al. , 2015. Exposure to endocrine-disrupting chemicals during pregnancy and weight at 7 Years of age: a multi-pollutant approach. Environ. Health Perspect. 123, 1030–1037. 10.1289/ehp.1409049. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Ahsan H, Perrin M, Rahman A, Parvez F, Stute M, Zheng Y, et al. , 2000. Associations between drinking water and urinary arsenic levels and skin lesions in Bangladesh. J. Occup. Environ. Med. 42, 1195–1201. [DOI] [PubMed] [Google Scholar]
  3. Bai J, 1997. Estimation of a change point in multiple regression models. Rev. Econ. Stat. 79, 551–563. 10.1162/003465397557132. [DOI] [Google Scholar]
  4. Barnett AG, van der Pols JC, Dobson AJ, 2004. Regression to the mean: what it is and how to deal with it. Int. J. Epidemiol. 34, 215–220. 10.1093/ije/dyh299. [DOI] [PubMed] [Google Scholar]
  5. Bates D, Mächler M, Bolker B, Walker S, 2015. Fitting linear mixed-effects models using lme4. J. Stat. Softw. 67, 1–48. 10.18637/jss.v067.i01. [DOI] [Google Scholar]
  6. Bhargava SK, Sachdev HS, Fall CHD, Osmond C, Lakshmy R, Barker DJP, et al. , 2004. Relation of serial changes in childhood body-mass index to impaired glucose tolerance in young adulthood. N. Engl. J. Med. 350, 865–875. 10.1056/NEJMoa035698. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Carignan CC, Cottingham KL, Jackson BP, Farzan SF, Gandolfi AJ, Punshon T, et al. , 2015. Estimated exposure to arsenic in breastfed and formula-fed infants in a United States Cohort. Environ. Health Perspect. 123, 500–506. 10.1289/ehp.1408789. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Carignan CC, Karagas MR, Punshon T, Gilbert-Diamond D, Cottingham KL, 2016. Contribution of breast milk and formula to arsenic exposure during the first year of life in a US prospective cohort. J. Expo. Sci. Environ. Epidemiol. 26, 452–457. 10.1038/jes.2015.69. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Chen CJ, Chiou HY, Chiang MH, Lin LJ, Tai TY, 1996. Dose-response relationship between ischemic heart disease mortality and long-term arsenic exposure. Arterioscler. Thromb. Vasc. Biol. 16, 504–510. [DOI] [PubMed] [Google Scholar]
  10. Claus Henn B, Ettinger AS, Hopkins MR, Jim R, Amarasiriwardena C, Christiani DC, et al. , 2016. Prenatal arsenic exposure and birth outcomes among a population residing near a mining-related superfund site. Environ. Health Perspect. 124, 1308–1315. 10.1289/ehp.1510070. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Concha G, Vogler G, Lezcano D, Nermell B, Vahter M, 1998. Exposure to inorganic arsenic metabolites during early human development. Toxicol. Sci. 44, 185–190. 10.1093/toxsci/44.2.185. [DOI] [PubMed] [Google Scholar]
  12. Eriksson JG, Forsen TJ, Osmond C, Barker DJP, 2003. Pathways of infant and childhood growth that lead to type 2 diabetes. Diabetes Care 26, 3006–3010. 10.2337/diacare.26.11.3006. [DOI] [PubMed] [Google Scholar]
  13. Eriksson JG, Kajantie E, Lampl M, Osmond C, 2015. Trajectories of body mass index amongst children who develop type 2 diabetes as adults. J. Intern. Med. 278, 219–226. 10.1111/joim.12354. [DOI] [PubMed] [Google Scholar]
  14. Fängström B, Moore S, Nermell B, Kuenstl L, Goessler W, Grandér M, et al. , 2008. Breast-feeding protects against arsenic exposure in Bangladeshi infants. Environ. Health Perspect. 116, 963–969. 10.1289/ehp.11094. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Fenton TR, Kim JH, 2013. A systematic review and meta-analysis to revise the Fenton growth chart for preterm infants. BMC Pediatr. 13, 59 10.1186/1471-2431-13-59. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Francesconi KA, Tanggaard R, McKenzie CJ, Goessler W, 2002. Arsenic metabolites in human urine after ingestion of an arsenosugar. Clin. Chem. 48, 92–101. [PubMed] [Google Scholar]
  17. Gardner RM, Kippler M, Tofail F, Bottai M, Hamadani J, Grandér M, et al. , 2013. Environmental exposure to metals and children’s growth to age 5 years: a prospective cohort study. Am. J. Epidemiol. 177, 1356–1367. 10.1093/aje/kws437. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Gebel TW, 2002. Arsenic methylation is a process of detoxification through accelerated excretion. Int. J. Hyg Environ. Health 205, 505–508. 10.1078/1438-4639-00177. [DOI] [PubMed] [Google Scholar]
  19. Gilbert-Diamond D, Cottingham KL, Gruber JF, Punshon T, Sayarath V, Gandolfi AJ, et al. , 2011. Rice consumption contributes to arsenic exposure in US women. Proc. Natl. Acad. Sci. U.S.A. 108, 20656–20660. 10.1073/pnas.1109127108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Gilbert-Diamond D, Emond JA, Baker ER, Korrick SA, Karagas MR, 2016. Relation between in utero arsenic exposure and birth outcomes in a cohort of mothers and their newborns from New Hampshire. Environ. Health Perspect. 124, 1299–1307. 10.1289/ehp.1510065. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Glymour MM, Weuve J, Berkman LF, Kawachi I, Robins JM, 2005. When is baseline adjustment useful in analyses of change? An example with education and cognitive change. Am. J. Epidemiol. 162, 267–278. 10.1093/aje/kwi187. [DOI] [PubMed] [Google Scholar]
  22. Groenwold RHH, White IR, Donders ART, Carpenter JR, Altman DG, Moons KGM, 2012. Missing covariate data in clinical research: when and when not to use the missing-indicator method for analysis. CMAJ (Can. Med. Assoc. J.) 184, 1265–1269. 10.1503/cmaj.110977. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Han E, Abrams B, Sridhar S, Xu F, Hedderson M, 2016. Validity of self-reported pre-pregnancy weight and body mass index classification in an integrated health care delivery system. Paediatr. Perinat. Epidemiol. 30, 314–319. 10.1111/ppe.12286. [DOI] [PubMed] [Google Scholar]
  24. Han Z, Mulla S, Beyene J, Liao G, McDonald SD, Knowledge Synthesis Group, 2011. Maternal underweight and the risk of preterm birth and low birth weight: a systematic review and meta-analyses. Int. J. Epidemiol. 40, 65–101. 10.1093/ije/dyq195. [DOI] [PubMed] [Google Scholar]
  25. Haque R, Mazumder DNG, Samanta S, Ghosh N, Kalman D, Smith MM, et al. , 2003. Arsenic in drinking water and skin lesions: dose-response data from West Bengal, India. Epidemiology 14, 174–182. 10.1097/01.EDE.0000040361.55051.54. [DOI] [PubMed] [Google Scholar]
  26. Hopenhayn-Rich C, Biggs ML, Fuchs A, Bergoglio R, Tello EE, Nicolli H, et al. , 1996. Bladder cancer mortality associated with arsenic in drinking water in Argentina. Epidemiology 7, 117–124. [DOI] [PubMed] [Google Scholar]
  27. Hopenhayn-Rich C, Biggs ML, Smith AH, 1998. Lung and kidney cancer mortality associated with arsenic in drinking water in Córdoba, Argentina. Int. J. Epidemiol. 27, 561–569. [DOI] [PubMed] [Google Scholar]
  28. Hopenhayn C, Ferreccio C, Browning SR, Huang B, Peralta C, Gibb H, et al. , 2003. Arsenic exposure from drinking water and birth weight. Epidemiology 14, 593–602. 10.1097/01.ede.0000072104.65240.69. [DOI] [PubMed] [Google Scholar]
  29. Hughes MF, 2002. Arsenic toxicity and potential mechanisms of action. Toxicol. Lett. 133, 1–16. [DOI] [PubMed] [Google Scholar]
  30. James KA, Byers T, Hokanson JE, Meliker JR, Zerbe GO, Marshall JA, 2015. Association between lifetime exposure to inorganic arsenic in drinking water and coronary heart disease in Colorado residents. Environ. Health Perspect. 123, 128–134. 10.1289/ehp.1307839. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Karagas MR, Tosteson TD, Blum J, Morris JS, Baron JA, Klaue B, 1998. Design of an epidemiologic study of drinking water arsenic exposure and skin and bladder cancer risk in a U.S. population. Environ. Health Perspect. 106 (Suppl. 4), 1047–1050. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Kwok RK, Kaufmann RB, Jakariya M, 2006. Arsenic in drinking-water and reproductive health outcomes: a study of participants in the Bangladesh Integrated Nutrition Programme. J. Health Popul. Nutr. 24, 190–205. [PubMed] [Google Scholar]
  33. Lai MS, Hsueh YM, Chen CJ, Shyu MP, Chen SY, Kuo TL, et al. , 1994. Ingested inorganic arsenic and prevalence of diabetes mellitus. Am. J. Epidemiol. 139, 484–492. [DOI] [PubMed] [Google Scholar]
  34. Laine JE, Bailey KA, Rubio-Andrade M, Olshan AF, Smeester L, Drobná Z, et al. , 2015. Maternal arsenic exposure, arsenic methylation efficiency, and birth outcomes in the Biomarkers of Exposure to ARsenic (BEAR) pregnancy cohort in Mexico. Environ. Health Perspect. 123, 186–192. 10.1289/ehp.1307476. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Larsen EH, Pritzl G, Hansen SH, 1993. Speciation of eight arsenic compounds in human urine by high-performance liquid chromatography with inductively coupled plasma mass spectrometric detection using antimonate for internal chromatographic standardization. J. Anal. At. Spectrom. 8, 557–563. 10.1039/ja9930800557. [DOI] [Google Scholar]
  36. Lindberg A-L, Goessler W, Gurzau E, Koppova K, Rudnai P, Kumar R, et al. , 2006. Arsenic exposure in Hungary, Romania and Slovakia. J. Environ. Monit. 8, 203–208. 10.1039/b513206a. [DOI] [PubMed] [Google Scholar]
  37. Mandal Badal Kumar, Ogra Yasumitsu, Suzuki KT, 2001. Identification of dimethyllarsinous and monomethylarsonous acids in human urine of the arsenic-affected areas in West Bengal, India. Chem. Res. Toxicol. 14, 371–378. 10.1021/TX000246H. [DOI] [PubMed] [Google Scholar]
  38. Meharg AA, Rahman MM, 2003. Arsenic contamination of Bangladesh paddy field soils: implications for rice contribution to arsenic consumption. Environ. Sci. Technol. 37, 229–234. [DOI] [PubMed] [Google Scholar]
  39. Moore MM, Harrington-Brock K, Doerr CL, 1997. Relative genotoxic potency of arsenic and its methylated metabolites. Mutat. Res. 386, 279–290. [DOI] [PubMed] [Google Scholar]
  40. Myers SL, Lobdell DT, Liu Z, Xia Y, Ren H, Li Y, et al. , 2010. Maternal drinking water arsenic exposure and perinatal outcomes in inner Mongolia, China. J. Epidemiol. Community Health 64, 325–329. 10.1136/jech.2008.084392. [DOI] [PubMed] [Google Scholar]
  41. Navas-Acien A, Umans JG, Howard BV, Goessler W, Francesconi KA, Crainiceanu CM, et al. , 2009. Urine arsenic concentrations and species excretion patterns in American Indian communities over a 10-year period: the strong heart study. Environ. Health Perspect. 117, 1428–1433. 10.1289/ehp.0800509. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Punshon T, Davis MA, Marsit CJ, Theiler SK, Baker ER, Jackson BP, et al. , 2015. Placental arsenic concentrations in relation to both maternal and infant biomarkers of exposure in a US cohort. J. Expo. Sci. Environ. Epidemiol. 25, 599–603. 10.1038/jes.2015.16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Rahman A, Vahter M, Smith AH, Nermell B, Yunus M, El Arifeen S, et al. , 2009. Arsenic exposure during pregnancy and size at birth: a prospective cohort study in Bangladesh. Am. J. Epidemiol. 169, 304–312. 10.1093/aje/kwn332. [DOI] [PubMed] [Google Scholar]
  44. Raml R, Goessler W, Traar P, Ochi T, Francesconi KA, 2005. Novel thioarsenic metabolites in human urine after ingestion of an arsenosugar, 2’,3’-dihydroxypropyl 5-Deoxy-5-dimethylarsinoyl-β- d -riboside. Chem. Res. Toxicol. 18, 1444–1450. 10.1021/tx050111h. [DOI] [PubMed] [Google Scholar]
  45. Rasmussen KM, Yatkine ALE, 2009. Weight gain during pregnancy. Reexamining the Guidelines 10.17226/12584. [DOI] [PubMed] [Google Scholar]
  46. Rivera-Núñez Z, Meliker JR, Meeker JD, Slotnick MJ, Nriagu JO, 2012. Urinary arsenic species, toenail arsenic, and arsenic intake estimates in a Michigan population with low levels of arsenic in drinking water. J. Expo. Sci. Environ. Epidemiol. 22, 182–190. 10.1038/jes.2011.27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Saha KK, Engström A, Hamadani JD, Tofail F, Rasmussen KM, Vahter M, 2012. Pre- and postnatal arsenic exposure and body size to 2 years of age: a cohort study in rural Bangladesh. Environ. Health Perspect. 120, 1208–1214. 10.1289/ehp.1003378. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Smedley PL, Kinniburgh DG, 2002. A review of the source, behaviour and distribution of arsenic in natural waters. Appl. Geochem. 17, 517–568. 10.1016/S0883-2927(02)00018-5. [DOI] [Google Scholar]
  49. Smith AH, Goycolea M, Haque R, Biggs ML, 1998. Marked increase in bladder and lung cancer mortality in a region of Northern Chile due to arsenic in drinking water. Am. J. Epidemiol. 147, 660–669. [DOI] [PubMed] [Google Scholar]
  50. Swanson J, Entringer S, Buss C, Wadhwa P, 2009. Developmental origins of health and disease: environmental exposures. Semin. Reprod. Med. 27, 391–402. 10.1055/s-0029-1237427. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Taylor VF, Li Z, Sayarath V, Palys TJ, Morse KR, Scholz-Bright RA, et al. , 2017. Distinct arsenic metabolites following seaweed consumption in humans. Sci. Rep. 7, 3920 10.1038/s41598-017-03883-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Tseng C-H, Chong C-K, Tseng C-P, Hsueh Y-M, Chiou H-Y, Tseng C-C, et al. , 2003. Long-term arsenic exposure and ischemic heart disease in arseniasis-hyperendemic villages in Taiwan. Toxicol. Lett. 137, 15–21. [DOI] [PubMed] [Google Scholar]
  53. Tseng CH, Tai TY, Chong CK, Tseng CP, Lai MS, Lin BJ, et al. , 2000. Long-term arsenic exposure and incidence of non-insulin-dependent diabetes mellitus: a cohort study in arseniasis-hyperendemic villages in Taiwan. Environ. Health Perspect. 108, 847–851. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Vahter M, 2009. Effects of arsenic on maternal and fetal health. Annu. Rev. Nutr. 29, 381–399. 10.1146/annurev-nutr-080508-141102. [DOI] [PubMed] [Google Scholar]
  55. Vahter M, 2002. Mechanisms of arsenic biotransformation. Toxicology 181–182, 211–217. 10.1016/S0300-483X(02)00285-8. [DOI] [PubMed] [Google Scholar]
  56. Valenzuela OL, Borja-Aburto VH, Garcia-Vargas GG, Cruz-Gonzalez MB, GarciaMontalvo EA, Calderon-Aranda ES, et al. , 2005. Urinary trivalent methylated arsenic species in a population chronically exposed to inorganic arsenic. Environ. Health Perspect. 113, 250–254. 10.1289/EHP.7519. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Watanabe T, Hirano S, 2013. Metabolism of arsenic and its toxicological relevance. Arch. Toxicol. 87, 969–979. 10.1007/s00204-012-0904-5. [DOI] [PubMed] [Google Scholar]
  58. WHO, 2018. Arsenic Fact Sheet. http://www.who.int/news-room/fact-sheets/detail/arsenic.
  59. WHO, 2000. Obesity: Preventing and Managing the Global Epidemic. Report on a WHO Consultation on Obesity. Geneva, pp. 3–5 June, 1997. [PubMed] [Google Scholar]
  60. WHO Multicentre Growth Reference Study Group, 2006. Reliability of anthropometric measurements in the WHO Multicentre growth reference study. Acta Paediatr. Suppl. 450, 38–46. [DOI] [PubMed] [Google Scholar]
  61. Winterbottom EF, Koestler DC, Fei DL, Wika E, Capobianco AJ, Marsit CJ, et al. , 2017. The aquaglyceroporin AQP9 contributes to the sex-specific effects of in utero arsenic exposure on placental gene expression. Environ. Health 16, 59 10.1186/s12940-017-0267-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Wood AJ, Raynes-Greenow CH, Carberry AE, Jeffery HE, 2013. Neonatal length inaccuracies in clinical practice and related percentile discrepancies detected by a simple length-board. J. Paediatr. Child Health 49, 199–203. 10.1111/jpc.12119. [DOI] [PubMed] [Google Scholar]
  63. Xu W-X, Liu Y, Liu S-Z, Zhang Y, Qiao G-F, Yan J, 2014. Arsenic trioxide exerts a double effect on osteoblast growth in vitro. Environ. Toxicol. Pharmacol. 38, 412–419. 10.1016/j.etap.2014.07.010. [DOI] [PubMed] [Google Scholar]
  64. Yang C-Y, Chang C-C, Tsai S-S, Chuang H-Y, Ho C-K, Wu T-N, 2003. Arsenic in drinking water and adverse pregnancy outcome in an arseniasis-endemic area in northeastern Taiwan. Environ. Res. 91, 29–34. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

supplementMtrl
NIHMS1598299-supplement-supplementMtrl.docx (802.3KB, docx)

RESOURCES